Explanations of adolescent risk taking invariably make reference to adolescents' beliefs in their own invulnerability to harm, and to the notion that teenagers' perceptions of risk are more biased than those of adults. Although the implications of these assertions are critical for theory, intervention and policy, little research has been conducted on adolescent risk perception. Research on adults suggests sources of bias, such as experience and perceptions of controllability, that may also influence adolescent risk perception, but only a few comparative studies have been carried out, and none have included a prospective analysis of the relationship of risk perceptions to behavior. The proposed longitudinal study will use a cohort-sequential design to examine risk perception during early, middle and late adolescence, and adulthood. Study aims are to identify systematic biases in adolescents' estimation of risks; determine whether adolescents and adults differ qualitatively and/or quantitatively in their risk perception biases; identify periods during adolescence in which differences between adolescent and adult risk perception are greatest; and examine the prospective relationship of risk perceptions to risk behavior over time. Three ethnically-diverse cohorts of male and female adolescents from grades 6, 8, and 10 (n=450) will be recruited using public school enrollment lists. After collection of baseline data, teens will be followed every six months for an additional three years, until grades 9, 11, and one year post high school. A sample of 25-30 years old adults (N=200), comparable in sex race(ethnicity to the adolescent samples, will be recruited from 10 colleges and places of employment. The adults will be assessed annually. Assessments will occur at community locations and will include measures on risk perceptions, negative experiences, perceived controllability, and behavior across a range of areas including sexual behavior, substance use, vehicle use, safety practices and personal health habits. In addition to providing knowledge of relevance to theories of health behavior, the results of the proposed study will provide specific knowledge about adolescent risk perceptions that is critical to the further development of effective, theory-based interventions to reduce risk taking in youth. The results will also have relevance to policy issues related to adolescent and informed consent.